/srv/irclogs.ubuntu.com/2016/10/14/#snappy.txt

mupBug #1571497 changed: snapd crashes on slot disconnection <Snappy:Fix Released by zyga> <https://launchpad.net/bugs/1571497>00:42
mupBug #1628592 changed: no camera slot in pi2 or pi3 image <Snappy:Fix Released by zyga> <https://launchpad.net/bugs/1628592>00:42
guy_davisHiya, anyone installed dockerd on snap core 16 beta?  Looking to evaluate Ubuntu Snappy as a minimal OS for hosting docker containers.01:02
mupBug #1633289 opened: network-observe plug is not working for dracnmap snap <Snappy:New> <https://launchpad.net/bugs/1633289>01:46
=== chihchun_afk is now known as chihchun
dholbachgood morning07:03
ejatmorning07:08
=== chihchun is now known as chihchun_afk
mardyhow can I specify which Ubuntu release should be used by "stage-packages" and "build-packages"?07:44
=== chihchun_afk is now known as chihchun
vicamohi there, do any one know where to send a merge proposal for ubuntu snap-confine package?08:32
vicamowould like to propose a change to enable running snappy on Ubuntu M1008:32
Rick_Hello08:36
Rick_Does someone know the solution of No ssh keys found, at creating user profile setup08:38
mupPR snapd#2145 opened: cmd/snap: update remove command help <Created by fgimenez> <https://github.com/snapcore/snapd/pull/2145>08:56
ogra_mardy, i dont think you can (by default it uses the hosts sources.list ...)09:12
mardyogra_: OK. It would be nice if there was a way to specify that in snapcraft.yaml, so that one would know in which release a certain snap was built09:15
ogra_mardy, well, for now you can put it in the description or so09:19
mardyyep09:19
ogra_long term -> file a whishlist bug ;)09:19
ogra_though since you dont really need to use any debs i'*m not sure it is an option thats wanted by default09:20
ogra_(you can make your snapcraft.yaml complex and build all deps for source for example ... i.e. you wouldnt need to use any debs if you like)09:20
zygavicamo: hey09:29
zygavicamo: I'm the maintainer of snap-confine, please send me a pull request on github09:30
zygavicamo: can you tell me more about what you were working on?09:30
vicamozyga: I have to modify the apparmor rules for ubuntu for snap-confine package09:31
vicamoI guess that's not for the rest of the world?09:31
zygavicamo: can you tell me more about what you need to modify and why?09:32
vicamozyga: it's an open bug in https://bugs.launchpad.net/ubuntu/+source/snap-confine/+bug/163336709:33
mupBug #1633367: missing ptrace options needed by snap-confine <snappy> <systemd> <Canonical System Image:New> <snap-confine (Ubuntu):New> <https://launchpad.net/bugs/1633367>09:33
* zyga looks09:33
zygavicamo: can you tell me which kernel version are you using on m1009:35
vicamozyga: a pretty outdated 3.1009:35
ogra_could be worse ... could be 2.18 :P09:36
zygavicamo: does it include the patched apparmor stack?09:36
vicamozyga: yes, it has, but again, outdated 3.1 or so09:36
zygavicamo: I mean the updated patch for 3.10 kernels?09:36
zygavicamo: one sec09:37
zygavicamo: https://github.com/snapcore/sample-kernels09:37
mupPR snapcraft#869 opened: Bring docs/upload-your-snap.md in line with http://snapcraft.io/docs/… <Created by dholbach> <https://github.com/snapcore/snapcraft/pull/869>09:37
zygavicamo: in any case, please fork snap-cofnine and open a pull request09:37
zygavicamo: the code is on github.com/snapcore/snap-confine09:37
zygavicamo: I'll work with jamie to review it09:38
zygavicamo: and be sure to try current master09:38
vicamozyga: that sameple kernel has a 3.10 tree, but its apparmor part is not up-to-date, either09:38
zygavicamo: a large change just landed09:38
zygavicamo: ah, I see, I wasn't aware of that09:38
Rick_At the first boot of snappy core, at the login/registration, where i must fill in my email, it is saying that there are no ssh keys found and that my account could not be breate09:38
Rick_created09:39
zygaRick_: your keys are taken from your launchpad.net account09:39
Rick_Could someone help me?09:39
vicamoI've also checked the rest of snap-confine code base on github, I need to land a change in the apparmor rules, not snap-confine itself09:39
zygavicamo: that is in src/snap-confine.apparmor.in09:40
Rick_Zyga: Thank you!09:40
vicamozyga: ah!09:40
Rick_Zyga: How can i get my keys?09:50
Rick_I created my account, and tried it again, but the same error, the account could not be created, no ssh keys found09:50
Rick_Aah, i must register a key, using launchpad.net i saw, gonna try i09:52
Rick_*it09:52
Brad__join09:55
Brad__ hello09:55
Brad__Anyone creating snaps?09:56
Brad__Anyone here?09:56
ogra_only 302 people09:56
Brad__Oh !09:57
davmor2ogra_: I'm not I'm breaking them :P09:57
ogra_heh, true :)09:57
=== Guest82060 is now known as dch
Brad__That's about as far as i have got, too09:58
ogra_thats a good start ;)09:58
zygaRick_: look at your launchpad profile, there's a SSH section there09:59
Rick_Thanks Zyga, i followed the instructions and it is working. :) Jeahh Thankss, This awnser was the only one i could find09:59
zygaRick_: I'm glad to hear that :)10:01
vicamozyga: https://github.com/snapcore/snap-confine/pull/170 pr ready, but it seems CI is currently broken?10:01
mupPR snap-confine#170: Add apparmor ptrace options for snap-confine <Created by vicamo> <https://github.com/snapcore/snap-confine/pull/170>10:01
Brad__Can anyone help me create a basic calculator, chat or notepad snap? I tried the yaml on github but it says an error msg. Do i need to install packages? Has anyone got some yaml code which will work?10:05
Brad__bump10:08
Brad__only 302 people10:13
ogra_!patience10:14
ubottuDon't feel ignored and repeat your question quickly; if nobody knows your answer, nobody will answer you. While you wait, try searching https://help.ubuntu.com or http://ubuntuforums.org or http://askubuntu.com/10:14
ogra_;)10:14
Brad__oh, thank :)10:14
ogra_it might be helpful if you pasted the error you get on paste.ubuntu.com and give the pastebin url so people can look at it10:15
Brad__okay, great!10:16
Brad__http://paste.ubuntu.com/23322759/10:31
ogra_Brad__, so you are using a PPA that has no xenial packages it seems10:32
Brad__okay -do you know how i can the xenial packages? Or exclude the PPA?10:34
ogra_snapcraft uses your hosts apt sources ... disable it on your PC and snapcraft wont try to use it10:34
ogra_(alternatively call "snapcraft cleanbuild" that will build inside a container and not use the info from the host machine)10:37
Brad__okay, thanks so much. do you know how i can disable it? :)10:37
ogra_well, how did you enable it ... ? :)10:37
ogra_you obviously set it up once ... to install ubuntu-tweak i guess10:37
Brad__oh, :/ :)10:38
ogra_just go to your software sources in the settings and turn it off there10:38
Brad__great. thank you.10:38
mupPR snapd#2146 opened: asserts: remove serial-proof type <Created by emgee> <https://github.com/snapcore/snapd/pull/2146>10:45
mectorssudo snap create-user --sudoer --json tadhackuser@gmail.com11:13
mectorserror: while creating user: cannot create user: device already managed11:13
mectors???11:13
mectorsWhy can I not add another user to Ubuntu Core 1611:13
mectorsfound it. You have to logout11:15
mattywhey folks, which repository should snap be coming from on trusty?11:20
mattyw(snapd that is)11:20
mupPR snapd#2147 opened: asserts: validate optional account username <Created by emgee> <https://github.com/snapcore/snapd/pull/2147>11:22
zygamattyw: eventually the same one11:22
zygamattyw: the trusty archive11:23
mattywzyga, I was expecting to just be able to do apt-get install snapd, but it can't be found11:23
joc_zyga: hi, if you are happy with the comments from jamie could you merge my PR in to your branch https://github.com/zyga/snapd/pull/211:25
mupPR zyga/snapd#2: Make sure gpio exported before apparmor setup <Created by jocave> <https://github.com/zyga/snapd/pull/2>11:25
zygamattyw: It's not available yet11:26
mupBug #1632453 changed: snapd won't start on dragonboard <Snappy:New> <snapd (Ubuntu):New> <https://launchpad.net/bugs/1632453>11:26
zygajoc_: I'll check that out next week, I'm on holidays, in theory, this week11:26
mupPR snapd#2146 closed: asserts: remove unused serial-proof type <Created by emgee> <Merged by pedronis> <https://github.com/snapcore/snapd/pull/2146>11:26
zygajoc_: (but hey, /media support landed today)11:26
joc_zyga: ok that's fine, lets see if we can get it resolved next week11:27
zygajoc_: we'll work on the release next week11:27
zygajoc_: face to face will be faster11:27
joc_+1 :)11:27
zyga(hopefully over beer at the bar)11:27
mattywzyga, ok thanks11:28
mattywzyga, is there a ppa it can be got from though?11:30
liuxgogra_, now I am removing a snap, but it shows me "2016-10-14T11:54:33Z ERROR cannot remove snap file "lights", will retry in 3 mins: umount: /snap/lights/x1: not mounted". what shall I do next? I just want to remove the app.11:57
ogra_well, why is the snap not mounted ?11:58
ogra_did you manually unmount it ?11:58
liuxgogra_, it was mounted yesterday. today, I want to install a new version. Before that, I wanted to remove it first.11:58
liuxgogra_, No, I did not do that.11:58
liuxgogra_, using snap list, it shows "lights                            x1              devmode,disabled,broken"11:59
ogra_well, it looks like you tinkered with it manually somehow11:59
liuxgogra_, frankly, I did not do that.11:59
ogra_check snap changes and take a look at the offending change there with "snap change <number>"12:00
liuxgogra_, I quit the command by Ctrl+c. then http://paste.ubuntu.com/23323101/12:01
ogra_well, obviously take a look at 9212:01
liuxgogra_, what can I look at?12:02
ogra_the details of that change12:02
liuxgogra_, in the step, it just happened like the message I showed you.12:02
liuxgogra_, where can I find the detail of the change?12:03
ogra_<ogra_> check snap changes and take a look at the offending change there with "snap change <number>"12:03
liuxgogra_, http://paste.ubuntu.com/23323114/12:04
ogra_Error   2016-10-14T11:51:46Z  2016-10-14T11:54:24Z  Remove security profile for snap "lights" (x1)12:05
ogra_now, no idea why that fails12:05
ogra_probably ask someone from the snapd team or security how o debug this further12:06
ogra_*how to12:06
ogra_(and file a bug if you are sure you havent manually tinkered with the snap in any way)12:06
liuxgogra_, it is ok. thanks. Have a good night! I really do not know how to debug this kind of issue.12:07
ogra_that is why i point you to people that do know :)12:07
ogra_(i dont either and for me it is also still early afternoon btw ;) )12:07
liuxgogra_, many thanks. it is a night here :) I may find someone to help out.12:09
ogra_perhaps zyga has an idea (seems mvo is off today)12:09
didrockseven if there has been some manual tinkering, it's a good use case to make snapd reliable against that and recover its state from it :)12:12
didrockslike not being mounted sounds like the code isn't there anymore, and it's safe to proceed12:12
ogra_yes, which is why i suggested to file a bug :)12:12
didrocksyep!12:12
=== chihchun is now known as chihchun_afk
kgunncjwatson: if you're about, i have a particular snap build failure that seems "sticky" to a particular project13:25
kgunnhttps://launchpadlibrarian.net/289526453/buildlog_snap_ubuntu_xenial_amd64_miral-snap_BUILDING.txt.gz13:25
kgunni mean i can build that locally & the error is a unknown to me...wondering if you could take a look and give me your $0.0213:26
ogra_kgunn, try adding the ubuntu image build PPA to your snap build13:29
ogra_we have a fixed bzr13:30
ogra_or enable proposed, the fix is there as well, but not yet landed13:30
kgunnogra_: ah thanks!13:30
kgunnogra_: so just proposed instead of "updates"13:30
ogra_bug 160620313:31
mupBug #1606203: Failed to build of snappy package on Launchpad: Invalid header value 'Basic U05BUEJVSUxELTE4NzAtMTQ2OTQyNjE0ODpjOTJkYzVjOWQ0OTg0ZGE5OWZlNGY1ZjI3ODRhMWJk\nOA==' <launchpad> <snappy> <verification-needed> <Bazaar:Fix Committed by vila> <Snapcraft:Invalid by sergiusens> <bzr (Ubuntu):Fix13:31
mupReleased> <bzr (Ubuntu Xenial):Fix Committed by cjwatson> <https://launchpad.net/bugs/1606203>13:31
ogra_kgunn, yeah13:31
kgunnta13:31
cjwatsonoh, I've been meaning to verify that13:36
cjwatsonkgunn: hey, if you enable proposed and it works, could you do me a favour and replace the verification-needed tag on that bug with verification-done?13:36
kgunnhappy to13:37
cjwatson'cos that was basically what I was going to test13:37
cjwatsonand definitely use proposed rather than the PPA, that way it's a valid test for SRU verification purposes13:37
kgunncjwatson: ah, same error...but wonder, is it b/c i'm targeting xenial?13:41
kgunnhttps://code.launchpad.net/~kgunn72/+snap/miral-snap/+build/693813:41
niemeyerjdstrand: ping13:44
cjwatsonkgunn: you didn't enable -proposed there13:44
jdstrandniemeyer: hey13:44
niemeyerjdstrand: Yo, good morning13:44
niemeyerjdstrand: How're things going in snap-declaration land?13:45
cjwatsonkgunn: and, well, yes it's because you're building from xenial, but the proper fix for that isn't to build from some other series, it's for us to get this fix verified :)13:45
cjwatsonkgunn: huh.  you did *try* to enable -proposed.  but no mention of it in the build log ...13:46
cjwatsonthat's weird13:46
jdstrandniemeyer: I have checks against the base declaration working, except for alternations. after that, will add the bits for when the store gives the payload13:46
kgunncjwatson: lemme know if you want me to try something else13:47
jdstrandniemeyer: so the review tools are pretty close. it took longer than I expected but it should be solid13:47
cjwatsonkgunn: let me scratch my head about this for a bit first13:47
kgunnyou bet13:47
jdstrandniemeyer: I don't know the status on the store side13:47
niemeyerjdstrand: As we discussed earlier in the week, the main concern at this point is the snap-declaration text itself13:47
niemeyerjdstrand: We need the text ready so that we can take out the hardcoded logic13:48
niemeyerjdstrand: We can import the initial declarations by hand if necessary13:48
niemeyerjdstrand: But we need them there so that we can update and release snapd13:48
jdstrandniemeyer: yes. my understanding was that the store guys were going to put that in the store db13:48
niemeyerjdstrand: Right, but do we have "that"? :)13:49
jdstrandroadmr: can you comment on that work ^13:49
niemeyerjdstrand: The question is really the snap-declaration body for the constraints13:49
roadmrjdstrand: sure, give me a bit as I'm OTP13:49
* jdstrand nods13:49
niemeyerjdstrand: We're in touch with matiasb as well, who's working on the feature and aware of these ideas13:50
jdstrandI see13:50
niemeyerjdstrand: So the main thing I'm bothering you about (sorry!) is the final/real text for the base and the individual snaps that allows us taking out the hardcoded logic.. with that in hands we can land the declarations and update snapd13:51
cjwatsonkgunn: ah, this is because you're building against a PPA and they have confusingly different rules13:51
matiasbjdstrand, fyi, from the store UI it is already possible as reviewer to push update plugs/slots for a snap declaration13:51
cjwatsonogra_: did you have a snap that builds only from the primary archive where you could test the bzr fix in -proposed?13:51
jdstrandmatiasb: ok. so, for example, I can handcraft the payload for lxd, and the store will regenerate the snap declaration for snapd to consume?13:52
ogra_cjwatson, hmm, not really ... all my snaps that need bzr actually also need to use the PPA for other bits13:52
matiasbjdstrand, exactly13:53
jdstrandok cool13:53
jdstrandniemeyer: so, since that is done ^ shall I transition to making the change to the base declaration in snapd? I need that anyway and I also worked through that a bit yesterday, so it should be quick13:54
jdstrandquickish13:54
niemeyerjdstrand: Yeah, that plus the few special allowances we have would be great to have ready, thank you!13:55
jdstrandit's possible I might need some help with the testsuite, but I'll know more when I get in there13:55
jdstrandok, I'll move to that13:56
roadmrjdstrand: finally of the phone, but yeah, what matiasb said, the store is already spitting out snap-declaration with plugs/slots which can be edited by reviewers. Plus it gets a small amount of validation against the most basic shot-in-the-foot scenarios14:02
jdstrandroadmr: ack. I made several commits to the review tools. I need to pivot to snapd now, but the tools should be ready for a pull pretty soon after I finish with snapd14:06
roadmrjdstrand: awesome! we can push that to the store when available, and there's a switch we can flip if things misbehave, so at least we'd call the tools "the old way" (no slots/plugs for preapproval). Recall we'll be sprinting next week so that may affect response times a bit14:08
jdstrandI definitely remember we'll be sprinting next week :)14:09
cjwatsonogra_: all right, let me see if I can hack something up14:10
ogra_i can surel dpo something like that too, but a nnit later today if you dont get to it14:10
roadmrjdstrand: hehe :)14:10
ogra_*but a bit14:11
cjwatsonogra_,kgunn: OK, verified now, I expect it'll be released on Monday (no SRU releases on Friday, normally)14:18
mupPR snapd#2140 closed: asserts,snap: validate attributes to a JSON-compatible type subset <Created by pedronis> <Merged by niemeyer> <https://github.com/snapcore/snapd/pull/2140>14:24
mupPR snapd#2138 closed: many: removed frameworks target, and some leftover frameworks helpers… <Created by chipaca> <Merged by niemeyer> <https://github.com/snapcore/snapd/pull/2138>14:39
mupPR snapd#2145 closed: cmd/snap: update remove command help <Created by fgimenez> <Merged by niemeyer> <https://github.com/snapcore/snapd/pull/2145>14:41
mupPR snapd#2107 closed: client,daemon,cmd: add payment-declined error kind <Created by pete-woods> <Merged by niemeyer> <https://github.com/snapcore/snapd/pull/2107>14:50
mupPR snapd#2139 closed: cmd/snap: tweak unknown command error message <Created by fgimenez> <Merged by niemeyer> <https://github.com/snapcore/snapd/pull/2139>14:51
kgunnogra_: hey what was the ppa containing the fix for this lp issue?15:08
abeatojdstrand, ofono PR refreshed after changing the comment. Spread tests are failing but it looks unrelated to the PR15:08
abeato"cannot bind mount /snap/core/current//var/lib/console-conf -> /var/lib/console-conf with flags 0x85007. errmsg: Permission denied"15:09
CroephaHi15:50
CroephaSo, using current version of snapcraft using kernel plugin is most current way to make a kernel snap for core 16 right?15:51
kyrofaCroepha, yep16:06
ogra_kgunn, https://launchpad.net/~snappy-dev/+archive/ubuntu/image16:06
Croephaand by current version i mean the version that ubuntu desktop 16.04 sais is stable?16:06
Croephasays16:06
Croephakyrofa: ok thanks :)16:07
jdstrandpedronis: hey, so I have an updated base declaration, but all the snowflake handling in the testsuite is taking me a bit to sort through16:18
pedronisjdstrand: we need a bit to coordinate16:19
pedronisjdstrand: do you want to push something for me to look at?16:21
pedronisjdstrand: also are now all interfaces mentioned at least once in slots or plugs?16:22
jdstrandlet me push something16:23
=== blaroche is now known as blaroche_
jdstrandpedronis: https://github.com/jdstrand/snapd/tree/finish-decl-based-checks16:25
pedronisjdstrand: so basically need to work to make test pass again?16:27
jdstrandpedronis: that and then to remove the hardcoded stuff for lxd, docker, etc, and then to make pass again.16:28
jdstrandpedronis: I saw a snowflake for snapd-control too16:28
pedronisyes16:28
pedroniswe didn't block it16:28
pedronisanyway before this can land16:29
pedroniswe need all the matching decl in the store16:29
jdstrandoh, it is deny-auto-connection: false still in the base declaration16:29
jdstrandso that is stilla  snowflake16:29
jdstrandpedronis: should I change that now?16:30
pedronisjdstrand: I think the idea is to change everything16:30
pedronisbut we need the decls16:30
pedronisthat one is snapweb?16:30
jdstrandI can supposedly add the decls16:31
pedronisand I don't know what else16:31
pedronisjdstrand: maybe matiasb can also help telling how to search things that use some of those plugs16:31
matiasbpedronis, jdstrand, we should be able to search CPI for plugs/slots, at least for plugs/slots declared in the yaml (I think, maybe cprov can confirm?)16:33
cprovmatiasb: I am not sure it's working, we never finished that implementation.16:35
matiasbah, ok, I can try and take a look to confirm16:36
jdstrandpedronis: ok, I updated snapd-control to not auto-connect but the testsuite doesn't show there is an error there, which I expected it would16:38
jdstrandpedronis: (I made the change in the base decl)16:39
Elleois there some way to make URL opening work from snappy apps? I'm currently trying to snap up podbird, but external links using Qt.openUrlExternally() just result in things like "Unable to detect a web browser to launch 'http://community.ubuntu.com/contribute/translations'"16:39
jdstrandmatiasb, cprov: I was told I would be able to add plugs and slots via the web form now. is that the case?16:40
pedronisjdstrand: that sound strange, wrong syntax, I see the test that should fail16:40
ogra_Elleo, try adding /usr/local/bin to your PATH ...16:40
ogra_Elleo, inside yor snap that is16:40
Elleoogra_: okay, thanks16:40
ogra_which reminds me ... jdstrand was there any conclusion about the /usr/local/bin stuff inside the core snap yet ?16:41
matiasbjdstrand, correct, when reviewing a snap revision, you should now have an interfaces section with a link to the form to update the snap declaration16:41
pedronisjdstrand: I need to have dinner, back in a bit16:41
jdstrandmatiasb: is that something I can do without an upload? ie, I go find lxd, then add the necessary plugs/slots and then it is done, or does it have to be tied to an upload?16:42
cprovmatiasb: https://search.apps.ubuntu.com/docs/#slots, not exactly a direct filter, more like 'given the provided slots, omit incompatible plugs' ...16:42
jdstrandcprov: I barely remember the context for that. I think I mentioned it to mvo and he agreed it needed to be fixed16:45
jdstranderr16:45
jdstrandcprov: nm16:45
jdstrandogra_: ^16:45
ogra_jdstrand, ah, well ... thats about teh xdg-open hack we ship (but also a fake "apt" that just echost "please use snap" )16:46
ogra_i guess a topic for next week then16:46
matiasbjdstrand, not tied to an upload, you can update the declaration at any time; link available there atm because it is expected that you may need to update a decl if review indicates so16:47
cprovjdstrand: It's definitely not useful atm and it's not even supported for snap-specific searches (v1/snaps/search), we should not count on this.16:49
cprovmatiasb: ^16:49
matiasbjdstrand, you can try with this search: https://search.apps.ubuntu.com/api/v1/snaps/search?fields=name,plugs,slots (4 pages)16:49
matiasbjdstrand, that will return all snap pakages, and their indexed values for plugs/slots, I guess that should help16:49
jdstrandmatiasb: in this url: https://search.apps.ubuntu.com/api/v1/package/Oulq6jI8qkI4ScWVafl3VsxSsd52HSGu is Oulq6jI8qkI4ScWVafl3VsxSsd52HSGu the snap id?16:53
jdstrandmatiasb: also, that page is incomplete16:54
jdstrandbut that's ok16:54
jdstrandtoday, all slots implementations are flagged for manual review by the store16:54
jdstrandactually, that's not true16:55
jdstrandI need everything16:55
cprovjdstrand: try this https://pastebin.canonical.com/168034/16:56
cprovfor overview purposes16:57
jdstrandcprov: that's incomplete. for example, locationd is not listed: https://myapps.developer.ubuntu.com/dev/click-apps/566617:01
cprovjdstrand: it's not published on stable channel17:02
jdstrandcprov: what is the query for alls snaps and all channels?17:02
cprovjdstrand: unfortunately not, search was designed to operate on stable only.17:03
jdstranddocker isn't either17:03
jdstrandI guess I'll just have to go with what I know is implementing the slot side and using the plugs17:04
jdstrandmatiasb: I'm looking at the web form. do I need to add 'slots:' to the slots textarea or will it add it for me?17:06
* jdstrand tries on a test snap17:07
cprovjdstrand: Snap admin/overview APIs are non-existent at moment and we are starting to miss them.17:09
jdstrandhmm17:14
jdstrandmatiasb: if I add this to slots: {"bluez": {"allow-connection": {"plug-publisher-id": [".*"]}}}17:14
jdstrandmatiasb: I get "could not sign assertion (cannot assemble assertion snap-declaration: plug-publisher-id in allow-connection in slot rule for interface "bluez" contains an invalid element: ".*")"17:14
jdstrandI can use something else, but that seems like a bug17:15
matiasbjdstrand, that's an error from SAS... pedronis, fyi ^17:16
jdstrand{"bluez": {"allow-connection": {"plug-snap-type": ["app"]}}} worked17:16
pedronisjdstrand: regexps are not supported in id lists17:16
pedronis(they are not in the spec as far as I know)17:16
jdstrandI thought they were supported anywhere a string is17:17
jdstrandanyway, no matter for now, I can do a different thing17:17
pedronisjdstrand: no17:17
pedronisjdstrand:  Plug and slot attributes are specified with the following syntax17:18
pedronisand the it mentions regexps17:18
jdstrandpedronis: so, I can add the snap declarations to the store. and I can pull out the lxd, docker, etc hard coded things. I think I'll need help with getting the testsuite working properly though17:18
pedronisjdstrand: understood17:18
jdstrandpedronis: I also don't know how to 'turn it on' per se. iiuc, the base declaration changes are what do that. is it enough for me to compile a snapd with my base decl change, then try to install something from the store and see that it fails, then adjust the snap decl in the store and see that it works?17:21
pedronisjdstrand: yes17:21
jdstrandok, let me try that17:21
pedroniswe still want to cover with the unit tests17:22
jdstrandof course17:22
pedronisbut yes, to try for real we need that17:22
jdstrandI'm not sure how to fake a snap declaration in the testsuite17:22
jdstrandiirc, we were testing primarily against the base decl, not a snap decl17:23
pedronisjdstrand: no we already did a bit of that, see the test for LXD17:24
pedronisand mockSnapDecl17:24
pedronisyou can pass in plugs and slots stanzas17:25
pedronis(not used yet but supported by that)17:25
jdstrandpedronis: ok, so I'm still not clear on who is doing what17:26
pedronisjdstrand: well it makes for you to fill in as much snap decl as possible, and then to push the branch as far as possible but reasonable and then I pick up from there17:28
pedronisjdstrand: I have another couple of things to finish/start first so I cannot simply jump on that right now17:28
jdstrandok. I think I see what I need to do with lxd. let's see how far I get17:28
jdstrandpedronis: fyi, https://github.com/snapcore/snapd/commit/173152497e8e3e9d1966982ca4cd875376d94590 - I expected the second test to fail17:48
pedronisjdstrand: why is the first one also passing, with IsNil?17:51
pedronisjdstrand: I agree, I need to try and debug17:51
jdstrandpedronis: well, on the first one I was trying to create a snap decl that would allow the lxd snap to auto connect17:52
jdstrandpedronis: did I not do it right?17:53
pedronisjdstrand: I really need to grab your branch and play a sec17:53
pedronisjdstrand: let me try17:55
pedronisjdstrand: you have lxd-support on both slots and plugs side of the base decl for example, remember only one side of those is used17:59
jdstrandpedronis: oh, the very first one shouldn't pass, you are right. I had IsNil, I meant to have NotNil17:59
jdstrandTestAutoConnectionLxdSupport17:59
jdstrandcand := s.connectCand(c, "lxd-support", "", "") <- should fail17:59
pedronisyea17:59
pedronisbut passes18:00
pedronisbecause of what I mentioned18:00
jdstrandlxdDecl := s.mockSnapDecl(c, "lxd", "J60k4JY0HppjwOjW8dZdYc8obXKxujRu", "canonical", plugsSlots) <- should pass18:00
jdstrandoh18:00
jdstrandhmm18:00
jdstrandok, that explains snapd-control too18:00
pedronisjdstrand: it's the rule we wrote down18:00
pedronisnot let me try to understand the other one18:01
pedronisnow18:01
jdstrandpedronis: remind me why only one side?18:01
pedronisjdstrand: you decided that18:02
pedronisor you mean technically?18:02
jdstrandI don't recall deciding that... maybe I was thinking of something else18:02
jdstrandanyway, let me see if I can make it work with the current implementation18:03
pedronisjdstrand:  # - the final decision is taken by evaluating the opinion of the snap plug,18:03
pedronis#   then the snap slot, then the base plug, and finally the base slot, in that18:03
pedronis#   order. The first of those that define any rule about the circumstance is18:03
pedronis#   taken as the final decision. No merging is performe18:03
pedronisd18:03
jdstrandthe no merging I was thinking was between the snap and base declarations18:04
jdstrandbut I see in the plugs side there is no auto-connection, so it goes through18:05
pedronisjdstrand: well, you also reviewed the code18:05
jdstrandman, this is tricky to get the base declaration correct18:05
pedronisit sort of means18:05
pedronisyou can only use one side for most stuff18:06
pedronisor it gets delicate18:06
jdstrandindeed18:06
jdstrandthat is the sort of thing that was hard to see until stuff was implemented and tried to be put into use18:06
jdstrandanyway, working through it18:06
pedronisI don't expect doing merge18:07
pedronisto make the life easier for the code18:07
pedronisso probably it wouldn't be easy to understand18:07
pedroniswhat happens18:07
pedronisthough it my DWIM18:07
jdstrandI'm not suggesting we change it. I'm only saying that it makes defining the base declaration tricky and I have to work through some of the corner cases18:07
pedronisjdstrand: yea, that's why being to write tests reasonably is good18:08
pedronisbeing able18:08
jdstrandhrm18:11
pedronisjdstrand: about there other test is not picking up the slots declaration, trying to see why18:12
jdstrandI'm not sure this is going to work right for super-privileged interfaces. maybe I need to rethink what that means18:13
jdstrandcause it makes no sense to use 'plugs' for a plugging snap cause you can't specify the plug-snap-id on the plugs side18:13
pedroniswell if you are in snap18:14
pedronisit doesn't make sense to limit by snap-id18:14
pedronisyou just say18:14
pedronisyes18:14
=== AndyWojo is now known as suburban_killade
jdstrandpedronis: I'm trying to be able to say "only the lxd snap may use lxd-support'18:15
pedronisjdstrand: you can say it on lxd snap18:15
pedronisthough18:15
pedronisanyway still working through your other failing test18:16
jdstrandif I use deny-connection: true in the slots side of the base declaration, what does that mean for an implicit slot?18:16
pedronis?18:16
pedronisit means nothing can connect18:16
pedronis(that's not specific to implicit though)18:16
pedronisbut you can override in the snap plug rule18:17
jdstrandpedronis: ok, let me put it this way: can you show me a base decl for implicit slot 'foo' that should be denied to everyone. and then a snap decl for a snap that plugs: [foo] that allows it?18:17
pedronisbase-decl:18:18
pedronis   slots:18:18
jdstrand(denied to everyone on the plugs side that is)18:18
pedronis   foo:18:18
pedronis   deny-connection: true18:18
pedronissnap-decl18:18
pedronisplugs:18:18
pedronis   foo:18:18
pedronis   allow-connection: true18:18
pedronis?18:18
jdstrandright, that is what I was planning to do, but what does that mean on the slots side for an implicit slot?18:19
pedronissorry18:19
pedronisI'm not parsing that sentence18:19
jdstrandie, will the core snap not connect?18:19
pedronisconnect to what?18:19
jdstrandthere are two side of a connection18:20
jdstranda slot and a plug18:20
pedronisjdstrand: we discussed this18:20
pedronisjdstrand: this checks happen at higher level18:20
pedronisso the implicit side does whatever18:20
pedronisit does18:20
pedronisthat's about installation18:20
pedronisI imagine18:20
pedronisjdstrand: we don't check these rules18:21
jdstrandso you are saying I should ignore the fact that an interface is implicit?18:21
pedronisuntil we are giving two sides18:21
pedroniss/giving/given/18:21
jdstrandand an implicit slot only has one side in the code?18:21
pedronisI don't know18:21
jdstrandright, me either18:21
jdstrandwhich was why I was trying to be careful here18:21
pedronisI suspect is lower level than this18:21
pedronisI haven't seen anything that connect those as things18:22
jdstrandlet me just try to work through it the way we were thinking (slot only in base decl and plug snap decl override)18:22
pedronisI suspect the implicit stuff is written when you have connections18:23
pedronisbut is not a connection18:23
pedronisor something like that18:23
jdstrandalso, what does this mean:18:23
jdstrandbase decl:18:23
jdstrandslots:18:24
jdstrand  foo:18:24
jdstrand    allow-installation: false18:24
jdstrandwithin the context of an implicit slot18:24
pedronisprobabbly means you cannot refresh the snap os18:24
jdstrandright18:24
jdstrandthat isn't great :)18:24
jdstrandwhich is why I had:18:24
jdstrandplugs:18:24
pedronisinstallation rules18:24
pedronisshould mostly be about18:25
jdstrand  lxd-support:18:25
pedronissnap-type18:25
jdstrand    allow-installation: false18:25
pedronisor be on the plugs side18:25
jdstrandbut that doesn't work18:25
pedroniswhy?18:25
jdstrandI mean, it would work, but then I have both sides in the base decl18:25
pedronisthat's ok18:25
pedronisas long they are not about the same kind of rule18:25
pedroniswe can add a sanity check test about that sort of stuff18:25
pedronisinstallation rules vs connection vs auto-connection18:26
pedronisare orthogonal18:26
pedronisthat order for each18:26
jdstrandthe allow-installation rules are weird though in that the snap id isn't in there18:26
pedronisis not about any rules in general18:26
pedroniss/that order/that order is/18:26
jdstrandI would have to do in the snap decl18:27
jdstrandplugs:18:27
jdstrand  lxd-support:18:27
jdstrand    allow-installation: true18:27
pedronisyes18:27
jdstrandI guess that is good enough18:27
jdstrandcause the snap id is up above18:27
pedronisyes18:27
jdstrandok18:27
pedronisI honestly  hope/expect18:27
pedronismost snap level18:27
pedronisrule to be simple turn on18:27
pedronissomething denied in general18:27
pedronisbut allowed for this one18:28
pedronissnap18:28
pedronisjdstrand: ok, I find out the issue with your tests as it is18:29
pedronisjdstrand: you are using  slot rules on the plug side18:32
pedroniseven the passing test is wrong18:32
jdstrandpedronis: I'm reworking this base on our conversation18:33
pedronisjdstrand: anyway let me show you the diff so we are on the same page18:33
pedronisjdstrand: http://pastebin.ubuntu.com/23324868/18:34
pedroniss/docker-support/lxd-support/18:34
pedronisand lxdDecl goes into the cand.SlotSnapDeclaration side18:35
jdstrandoh, haha18:35
pedronisit's lxd is offers the slot18:35
pedroniswe check slot rules only on the slot snap18:35
jdstrandlxd-support is an implicit slot18:35
pedronisand plug rules only on the plug snap18:35
pedronisah ok18:36
pedronisthen the test is a bit non-sensical18:36
pedronisbut anyway we discussed18:36
pedronisthis stuff18:36
jdstrandlet me work through it18:36
pedronisok18:36
pedronisbut the infra seems to do the right things18:36
pedronisbut positive tests are delicate18:37
pedronis(given that allow is the default)18:37
jdstrandwe also discussed it weeks ago. putting it into practice with the nuances that are there is tricky18:37
* jdstrand is working through it18:37
pedronisjdstrand: I understand, just saying not seeing a bug yet18:37
=== drizztbsd is now known as timothy
=== nacc_ is now known as nacc
=== souther_ is now known as souther
=== xnox_ is now known as xnox
=== om26er_ is now known as om26er
=== thomi_ is now known as thomi
=== thibautr__ is now known as thibautr_
=== dobey_ is now known as dobey
=== Trevinho_ is now known as Trevinho
=== benoitc_ is now known as benoitc
mupPR snapd#2148 opened: many: introduce an assertion format iteration concept, refuse to add unsupported assertion <Created by pedronis> <https://github.com/snapcore/snapd/pull/2148>19:50
=== cjwatson_ is now known as cjwatson
barryhi all; i'm wondering if the python plugin for snapcraft has changed has changed recently.  my snapcraft.yaml that used to build ubuntu-image now does't get all the stage-packages installed correctly: https://github.com/CanonicalLtd/ubuntu-image/blob/master/snapcraft.yaml19:59
barryit worked fine for 0.7ubuntu1 (2016-10-06)20:00
barrybut now it can't import 'attrs'20:00
sergiusensbarry define recently20:08
barrysergiusens: as of oct 620:08
sergiusensbarry the deb, right? Then the answer is no https://launchpad.net/ubuntu/+source/snapcraft20:09
sergiusensunless you updated much after that20:09
barrysergiusens: interesting.  i'm fairly sure i would have picked that one up20:10
sergiusensbarry we run an extra `pip wheel` since then and do an explicit `pip download`20:12
sergiusensbarry for which I wanted to follow up with you on some other topic; seems there is a change in direction on what `install_data` should do in setup.py and that the whole thing is moving into a declarative setup. Any insight on that? I want the default plugin to follow the future and if necessary come up with another plugin that would live in the past20:14
barrysergiusens: i see all the whls in parts/ubuntu-image/packages but not in stage/user/lib/python3/dist-packages.  do i need to explicitly stage things in stage-packages?20:14
sergiusensbarry those should be in lib/pythonXX/site-packages20:14
barrysergiusens: there's not even a site-packages directory in stage/usr/lib/python3.520:15
sergiusensbarry oh, add lib to your filters20:15
barry(or a dist-packages)20:15
barryah20:15
barrythanks, trying that20:15
barrysergiusens: do you mean setup.cfg?20:16
sergiusensbarry this was part of the warning email I sent a month ago or so; $SNAP/lib is PYTHONUSERBASE20:16
barrysergiusens: i feel a sense of deja vu ;)20:16
barrysergiusens: thanks20:17
sergiusensbarry np20:18
sergiusensbarry this thread https://github.com/pypa/pip/issues/2874#issuecomment-10942948920:18
barrysergiusens: i have a short day today.  let me look at this in detail on monday20:19
sergiusensbarry sure, no rush; it is future planning after all20:19
barrycool20:20
mupPR snapd#2149 opened: Finalize declaration based checks <Created by jdstrand> <https://github.com/snapcore/snapd/pull/2149>20:34
=== nsg__ is now known as nsg
=== verterok` is now known as verterok
jdstrandpedronis (cc niemeyer): fyi, https://github.com/snapcore/snapd/pull/2149. see the comments-- I updated the store21:13
mupPR snapd#2149: Finalize declaration based checks <Created by jdstrand> <https://github.com/snapcore/snapd/pull/2149>21:13
* jdstrand is heading out for a while21:13
niemeyerjdstrand: Sounds great, thank you!22:18
mupBug #1633638 opened: Can no longer install snaps with multiline plugs <Snappy:New> <https://launchpad.net/bugs/1633638>22:36
mupPR snapd#2150 opened: tests: unflake ubuntu-core-reboot <Created by niemeyer> <https://github.com/snapcore/snapd/pull/2150>22:39
niemeyerjdstrand, matiasb: this looks like a bug in the assertion text, or in the assertion parser:22:40
niemeyerhttps://bugs.launchpad.net/snappy/+bug/163363822:40
mupBug #1633638: Can no longer install snaps with multiline plugs <Snappy:New> <https://launchpad.net/bugs/1633638>22:40
pedronisniemeyer: doesn't make a lot of sense, the parsing should be the same on the server and client23:24
pedronismatiasb: did we ignore some error somewhere ? ^23:27
pedronisor it's roundtripping bug23:29
pedronisbut no23:32
niemeyerjdstrand: Huston, we have a problem23:51
niemeyerjdstrand: We need to revert the new decls in the store.. I failed to anticipate the fact old snapd's can't parse maps in assertions.. we'll need to tune the infra early next week23:54
niemeyerMeanwhile, can you please take out the new slot/plug syntax, and keep it safe so we can readd next week?23:55

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!